The Functional Biology of Strombidium Sulcatum, a Marine Oligotrich Ciliate (Ciliophora, Oligotrichina)

MARINE ECOLOGY PROGRESS SERIES l Vol. 48: 1-15, 1988 - Published September 5 Mar. Ecol. Prog. Ser. l I

The functional biology of Strombidium sulcatum, a marine oligotrich ciliate (Ciliophora, Oligotrichina)

Tom ~enchel',Per R. ~onsson~

' Marine Biological Laboratory (University of Copenhagen). Strandpromenaden 5, DK-3000 Helsingar, Denmark Tjarno Marine Biological Laboratory, PI. 2781. S-452 00 Striimstad, Sweden

ABSTRACT: The bioenergetics and swimming behaviour of the marine oligotrich ciliate Strombidium sulcatum were studied. It grows on bacteria and heterotrophic flagellates with a maximum rate around 0.1 h-'. Cell yield is 3.3 X 10-5 ciliates bacterium-' and 7.7 X IO-~ciliates flagellate-' indicating a gross growth efficiency around 0.5. Starvation of S. sulcatum is accompanied by a 10-fold reduction in cell volume with a mean survival time of about 70 h. Clearance decreases rapidly for particles smaller than about 2 pm and cannot be detected for 0.21 pm beads. Maximum clearing rate for 2.83 pm latex beads is 1.6 $1 h-', close to the clearing rates found for bacteria, (1.5 p1 h-') and flagellates (2 $1 h-') and also to the calculated volume of water filtered (2 5 111 h-'). S. sulcatum efficiently accumulates in patches of bacteria. Analysis of ciliate behaviour revealed that this is caused by changes of swimming pattern from a helical path to a closed circle in response to bacterial density and that this response is chemically induced. This modulation of swimming behaviour when encountering a bacterial patch results in a 20- fold decrease in transportation rate with only a 40 % reduction in swimming speed. Video recordings of swimming ciliates in strobe-light suggest that the membranelles can change their position to affect the axial rotation of the ciliate. Absence of axial rotation results In a circular swimming path. In addition to this kinetic response, a transient response was observed which further enhances the probability of finding and remaining in bacterial patches. It was also found that S. sulcatum changes its vertical distribution in response to nutntional status; thls is at least in part due to the lunetic response, but change in cell morphology may also play a role. Findings are interpreted as adaptations for exploiting heterogeneous environments. The scale of the patchiness which the ciliate can exploit is about 1 m.

INTRODUCTION (Naitoh & Eckert 1974) and even gravity (Fenchel & Finlay 1984, 1986a). However, little is known about the Several studies recognize the sign~ficance of pro- ecological consequences of these sensory capabilities tozoa in aquatic ecosystems (e.g. Fenchel 1987) and and only a few studles on these exist (Fenchel & Finlay there is a growing interest in their functional properties 1984, Finlay & Fenchel 1986, Finlay et al. 1987). Given in relation to their ecological role. Recent work has the constraints of protist behaviour it is interesting to studied e.g. bioenergetics and food selection in some learn how they orient themselves in heterogeneous groups, but other aspects are poorly understood. An environments and with what efficiency they exploit important aspect of protozoan life on which information patchily distributed resources. Fenchel & Finlay (1984, is scarce and speculations abound concerns the ecolo- 1986b) and Finlay et al. (1987) have shown how seem- gical significance and mechanisms of behaviour. ingly complex distributions of Limnic ciliates can be Protozoa possess simple sensory organelles and a low explained by combinations of a few simple types of capacity for information processing and so are likely to behaviour. This indicates that simple mechanisms may have a limited repertoire of behaviours. Nevertheless, also be used by protozoa to exploit patchy food resour- physiological studies show that protozoa can respond ces. In this paper we describe the functional biology of to light (Song et al. 1980, Reisser & Hader 1984), a wide a marine oligotrich ciliate including an experimental range of chemicals (Van Houten et al. 1975, 1981, analysis of behaviour in prey gradients and its ecolo- Levandowsky & Hauser 1978), mechanical stimuli gical implications.

O Inter-Research/Printed in F. R. Germany Mar Ecol. Prog. Ser. 48: 1-15, 1988

MATERIALS AND METHODS 0.8 pm blue, 1.1 pm fluorescent and 2.83 pm non-dyed (Biosphere, Polyscience, USA). Four different concen- Materials. Strombidium sulcatum Claparede & Lach- trations for each particle size were used. After 5 min mann was isolated both from the Mediterranean Sea the experiments were stopped with glutaraldehyde (Ewer et al. 1985) and from Lmfjorden, Denmark, and (2.5 Oh) and the ciliates (10 to 30 cells) examined mi- subsequently cultured on mixed bacteria supplied with croscopically to count individual, ingested beads wheat grains. For observations of S. sulcatum with (SE< 10 %), and the volume cleared of particles was scanning electron microscopy, cells were fixed in a calculated. Experimental details, precautions and limi- mixture of Os04 (2 YO w/v) and HgC12 (sol. sat.), (6:1), tations of this technique are discussed in Fenchel filtered onto 1 pm nuclepore filters, dehydrated in a (1986). critical point dryer, covered with gold and examined Clearing rates of bacteria and heterotrophic with a Jeol scanning electron microscope. An ultra- flagellates were also calculated from growth experi- structural description of S. sulcatum can be found in ments. Clearing rate for ciliates growing on bacteria Faure-Fremiet & Ganier (1970). Cultures of the oligo- was calculated as: F,,, = bi,,,/(K.Y), where F,,, is trich ciliates S. reticulatum (Leegaard) and S, vestitum maximum clearing rate, p,,, is maximum growth rate, (Leegaard), originally isolated from Kosterfjorden K is the halfsaturation constant and Y is yield (Fenchel (Sweden), were fed a prasinophycean algae, 1982). In the experiments where Strombidium sul- Pyramimonas sp. In all experiments with bacteria a catum was grown on heterotrophic flagellates clearing Pseudomonas sp. grown on seawater agar medium rate was also calculated from the rate of disappearance (Zobell 1941) was used. of flagellates. Growth experiments. In separate batch cultures the Attraction of ciliates to patches of prey. Colonies of growth rate of Strombidium sulcatum was measured as bacteria were suspended in sterile seawater and a function of concentration of bacteria Pseudomonas immobilized with heat (80 "C). A drop of bacterial sus- sp., and of a heterotrophic flagellate, Pteridomonas pension stained with methylblue was injected into a danica. Cultures with bacteria were prepared by Sedgewick-Rafter counting chamber already filled scraping colonies from agar plates and suspending with 1 m1 of a suspension of Strombidium sulcatum. them in filtered and autoclaved seawater (Fenchel The accumulation of ciliates in the patch of bacteria as 1982). The flagellate was grown on Pseudomonas sp. a function of time was then followed wlth a dissecting in batch cultures and used when most bacteria had microscope. The diffusion of the bacterial patch (and of been ingested. All experiments were carried out at any leaking substances) could be monitored by the room temperature (20 "C). Changes in number of blue stain and covered 2 to 5 % of the total chamber ciliates and prey in the growth experiments were area. Similar experiments were also performed with 2 followed by sampling at regular time invervals. other Strombidium species, S. reticulatum and S. ves- Ciliates and heterotrophic flagellates were fixed in titurn. The accumulation rate of S. sulcatum was also Lugols solution and counted in a counting chamber. studied (by video-recording, see below) as a function of Bacteria were quantified spectrophotometrically at mean distance from the patch. This was done by vary- 660 nm and converted to cell number using standards ing the volume of ciliate culture in the counting of known concentrations. Growth rate was calculated chamber. from the initial phase of exponential growth. Cell yield Test of a chemical cue. We examined whether the in growth experiments was determined as the concen- attraction of Strombidlum sulcatum in the presence of tration of ciliates at the moment when cell volume bacteria was caused by chemosensory responses. starts to decrease divided by the inltial concentration Three capillary tubes were filled with a suspension of of prey. living bacteria, a peptone solution and filtered and To examine effects of starvation, survival and cell autoclaved seawater, respectively. The tubes were volume of Strombidium sulcatum was studied follow- sealed with petrol jelly in one end and with a cotton ing the reduction of food in the growth expenments. plug in the other. The cotton plug was firmly packed to Cell volume was calculated from linear dimensions (of retain the bacteria inside the tube while allowing dis- Lugol-fixed cells) assuming a geometric shape solved substances to leak out. The tubes were placed in described by 2 ellipsoids. a culture dish with filtered and autoclaved seawater Particle selection. Particle selection and ingestion and 50 washed S. sulcatum were added. After about 4 h rates of Strombidium sulcatum were determined in the accumulation of ciliates around the capillary tubes particle uptake experiments. Exponentially growing was noted. ciliates were incubated together with a series of 4 Analysis of swimming pattern. Swimming ciliates concentrations of suspended latex beads of the follow- placed in a Sedgewick-Rafter chamber were either ing kinds: 0.21 pm fluorescent, 0.51 Km fluorescent, photographed in dark field (1 s exposure) or recorded Fcnchel & Jonsson: F~~nct~onalb~ology of an ol~gotr~cllc~liiltr

with a video recorder (Sony video camera DXC-1O1P apparent geotaxis is the result of hydrodynamical and Panasonic recorder AG-6200) fitted to a dissecting reorientation due to gravitational sinking and an nlicroscope. Video-recorded movements of individual asymmetric cell shape. Density of the medium was ciliates were analyzed by plotting ciliates frame by increased with heavy water (deuteriumoxide, p = frame on a plashc film covering the monitor screen 1.11 g and the swimming angle to the vertical of allowing a time resolution of 2; S. Swimming S. sulcatum was determined from photographs taken behaviour as a function of prey density was studied through a horizontally set dissecting microscope with either around a patch of bacteria (prepared as 1 s exposure time. The long exposure time outlined the described above) or in a series of culture dishes with ciliates as white paths on the photographs, but since it different concentrations (104 to 108 ml-l) of homogene- was not possible to dishnguish the polarity of the paths ously suspended, living bacteria. To reduce bacteria only absolute values of angles (0 to 90") to the vertical coming from the stock cultures, ciliates were washed could be measured. To further test the hypothesis that several times in filtered and autoclaved seawater the non-randon1 vertical distributions were caused by before the experiments. hydrodynamical reorientation of cells, ciliates were The behavioural response of Strombidium sulcatum treated with NiN03 (1 mi\/l) which blocks ciliary rever- swimming in a chemical gradient without prey was sals and so random reorientation. studied in the vicinity of small agar blocks with peptone or bacterial extracts. An agar cube of about 0.5 mm3 was cut out and placed in a Sedgewick-Rafter RESULTS AND DISCUSSION counting chamber. Ciliates were then added and swimming paths were reconstructed as described above. Natural habitats and morphology Study of membranelle function. Strombidium sulcatum swimming in circles or in helices were video Strombidiliin sulcatum has previously been reported recorded through a microscope (200 to 1000~)using from marine sediments rich in organic material (Faure- sti-obe-light as light source. The strobe flashes were Fremiet 1924, Kahl 1932, Faure-Fremiet & Ganier sufficiently short to freeze the beating membranelles. 1970), from algal mats in a tidal marsh (Borror 1965)

The short strobe flash (< ''.,.l S) exposes the video and isolated from plankton (Fbvier et al. 1985). In the pickup device to the image for a brief fraction of the present study we have observed S. sulcatum on the field time and the pattern of resulting electilcal charge surface of organically rich sediments and on Beggiatoa is stored until it is swept at the standard scan rate. The mats. The observed habitats of S. sulcatum together use of a solid-state pickup device (CCD) ensured mini- with the behaviour described below indicate that it is mal lag, burn and geometric distortion (for more details primarily a benthic form which occasionally occurs in see Inoue 1986). Sequences were played frame by the plankton. Because it grows fast on bacteria and frame and membranelles traced from the screen on heterotrophic flagellates even a single individual in a plastic films. large pelagic water sample is likely to grow into a Vertical distribution. The vertical distribution of dense population upon enrichment or if the sample is Strombidium sulcatum as a function of starvation was left to stand. studied In wide 100 m1 test tubes. Tubes were placed in The general nlorphology of Strombidum sulcatum is a stirred water bath to reduce heat convection and similar to most other Stron~bidiumspecies with an open filled with a dense culture of growing ciliates and adoral zone of membranelles (14 + 7 oral membran- placed in the dark. The vertical distribution of the elles) consisting of closely set cilia, a conical shape and population was monitored over 1 wk taking 0.2 m1 the cap of cortical plates covering the posterior end samples at 4 depths from surface to bottom. In a sepa- (Fig. 1). It may be distinguished from most other rate experiment the change of vertical distribution species of Strombidium by the relatively small cap of upon feeding of starved ciliates was studied in a spec- cortical plates and its dimensions (length: 30 to 35 ,um, trophotometer cell. greatest diameter: 20 to 25 pm); moreover it is not To test if the ability of Strombidium sulcatum to flattened as many other benthic species. In contrast to actively swim downwards is caused by true geotaxis, many planktonic species which are green due to incor- implying physiological perception of gravity, an exper- porated plastids (Laval-Peuto & Febvre 1986, Jonsson iment was designed where swimming direction was 1987, Stoecker et al. 1987), S. sulcatum is colourless. studied as a function of density of the medium. The However, there is at least one planktonic species which hypothesis was that raising the medium density and is similar in appearance to S. sulcatum; it conforms to thus reducing the sinking rate of the ciliates should not the S. en~ergensdescribed by Leergaard (1915), but dramatically alter geotaxic behaviour if a gravity-sens- this specles does not show the changes in swimming ing organelle exists, but this would be the case if an pattern as described below. Mar. Ecol. Prog. Ser 48: 1-15, 1988

-U- II Fig 1. Strornbidium sulcaturn. Above: SEM micrographs; left:growing cell, nght: starving cell; Scale bar = 10 pm. Below: l S exposure photographs showing swlmrning paths; leA: growing cells, nght: moderately starving cells. White lines are spaced I., 1 mm apart Fenchel & Jonsson. Functional biology of an oligotrich ciliate

Bioenergetics catum responds with 1 or 2 final divisions without any intermediate cell growth resulting in a concomitant When a small number of Strombidium sulcatum are reduction of cell volume (Fig. 2). During the following inoculated together with Pseudomonas sp. or period there is a further gradual decrease in cell vol- Pteridomonas danica, ciliates start to divide and ume. In our experiments the response to the reduction increase exponentially in numbers (Fig. 2). The numer- of food was sudden and dramatic for ciliates growing ical response (Fig. 3) could be fitted to a hyperbolic on bacteria but less marked and gradual for ciliates function which has been found for other growing growing on flagellates; we do not know the reason for ciliates and heterotrophic flagellates (Heinbokel 1978, this. Survival is high during the first 50 h of starvation; Fenchel 1982, Jonsson 1986). Maximum growth rate is thereafter mortality increases and after 85 h only 10 % about 0.12 h-' (T2 5.8 h). Cell yield is 3.3 X 10-S of the population remains. ciliates bacterium-' and 7.7 X IO-~ciliates flagellate-' The size spectrum of ingested particles is shown in and nearly invariant with growth rate (Fig. 4). The Fig. 5. Clearance shows a marked decrease for 1.1 pm volume of the ciliate is about 9000 pm3 when growing latex beads relative to 2.8 pm beads and 0.51 ym beads on bacteria (volume: 0.6 pm3) and about 6500 pm3 are cleared at only 5 % of the rate shown for 2.83 pm when growing on flagellates (volume: 80 pm3); in both beads. No ingestion of 0.21 pm beads could be cases yield expressed as volume ciliates per volume detected. The reduction of clearing rate below 2 pm is ingested food is around 50 % assuming that fixation did consistent with the distances between membranelles not appreciably change cell volume. (Fig. 1) which are supposed to act as a filter (Fenchel When food supply is exhausted, Strombidiuni sul- 1986, Jonsson 1986). Strombidium species have about 7

a%yrber of cells

Fig 2. Strombidiurn sulcatum. (a) \ Batch culture initiated with 5 X 106 bacteria ml-' showing ciliate number and volume. (b) Batch culture initiated with 4 X 104 heterotrophic flagellates ml-' showing the number of ciliates and flagellates together with ciliate volume hours

- a 0.1 - fl m 0.05 m/!'

Fig.3. Strombidium sulcatum. Growth rate a as function of food concentration, (a) on bac- 024 6810 0 1 2 teria, (b) on flagellates bacteria ml-l x 106 flagellates ml-l x lo4 6 Mar. Ecol. Prog Ser. 48: 1-15. 1988

ming velocity of 600 pm s-' when feeding and a cross- sectional area including the membranelles of 1200 pm2. Volume-specific clearance is about 2 X 105 volumes h-' which is consistent with most studied polyphy- menophoran ciliates (Fenchel 1986). The maximum growth rates of S. sulcatum are similar to other small ciliates feeding on bacteria (Fenchel & Finlay 1983). kvier et al. (1985) presented estimates of growth rates of S. sulcatum growing on mixtures of marine bacteria. Results vary considerably between treatments (0.03 to 0.14 h-'), maybe because of varying composition of bacteria. Maximum growth rate of the planktonic S. retlculatum growing on micro-algae was reported to be more modest, 0.036 h-', (Jonsson 1986) but since the experiments were carried out at 12 "C growth rates might be doubled at 20 "C. Still, the estimate seems somewhat lower than for S. sulcatum perhaps reflect-

1 I 1 11111 101 / I 1 l 11111 ing differences in food quality. The gross growth effi- 0 .l 1 10 bacteria ml-l x107 (m) flagella!es ml-l x104 (0) ciency for S. sulcatum expressed as cell yield in terms of volume was consistent with the estimates between Fig. 4. Strombidium sulcatum. Cell yield as function of initial 40 to 60 % which have been found for several other food concentrations protozoa (Fenchel & Finlay 1983). Particle size selection of S. sulcatum is comparable to the planktonic species S. reticulatum and S. vestitum (Jonsson 1986) which showed a similar drop in clearing rate for particles smaller than 2 pm. Thls characteristic is shared by many polyhymenophoran ciliates that retain particles between the membranelles (Fenchel 1986). The low clearing rates for particles smaller than 1 pm indicate that high food concentrations are necessary if Strombidium sulcatum is to grow on planktonic bacteria. Assuming that the ability to modulate growth rate in relation to available food is an adaptation to changing food conditions, the growth half-saturation constant may be used as a rough measure of the average food concentration experienced by the organisms. Half-sat- - uration concentration with Pseudomonas sp. as food is diameter, pm 2.5 X 106 cells ml-' and with flagellates it is 6.5 X 103 Fig. 5. Strombidiurn sulcatum. Maximum clearing rate as cells ml-l. The concentration of flagellates corresponds function of particle size. Clearance for 2 83 pm particles is set well to average total nanoplankton of marine plankton to 100 % at least in coastal areas (Andersen & Snrensen 1986). This also applies to the concentration of bacteria but small oral membranelles which are more closely set the Pseudomonas sp. used in our experiments is a large (distance about 0.5 km) than the larger membranelles. rod-shaped bacterium measuring 1.5 X 0.8 pm while It may be speculated that the plateau between 1.1 and bacteria in the marine pelagial are generally smaller 0.8 pm beads is caused by retention of these particles and often spherical. McManus & Fuhrman (1986) by the small oral membranelles which become leaky to measured the size of marine planktonic bacteria and particles around 0.5 pm as indicated in Fig. 5. Thus the estimated an average diameter of 0.5 pm. This smaller filtering system may be considered as consisting of 2 bacterial size will result in lower clearing rate (conser- filters with different mesh sizes and flow rates. In abso- vatively by a factor of 5) and less food for each captured lute terms maximum clearing rate of 2.83 pm beads particle by S. sulcatum. To obtain the half-saturated was 1.6 ~11h-' which was close to the calculated clear- growth rate the concentration of natural bacteria would ance of bacteria and flagellates of about 1.7 and 2.2 p1 have to be around 7 X 10' ml-l, concentrations nor- h-' respectively. These estimates agree with a calcu- mally found only in sediments, along surfaces or locally lated filtering rate of about 2.5 yl h-' assuming a swim- around decaying organic matter. Fenchel 61Jonsson: Functional biology of an oligotrich ciliate 7

Fig. 6. Strombidium sulcatum. Accumu- lation in a patch of bacteria. The curve in (a) is based on the regression line of the same data as presented in (b). Areas in- and outside the patch were 30 and 970 mm2, respectively. Results can be explained as chemokinesis if the motility of the ciliates is about 50 times higher out- s~dethan inside the patch minutes minules

Behaviour

When Strombidium sulcatum cells swim in a flat chamber with a small patch of concentrated livlng or heat-killed bacteria, ciliates immediately accumulate (Fig. 6). In the filled chamber (2 X 5 cm) an equilibrium distribution is reached with ca 60 O/O of the ciliates swimming in the bacterial patch (ca 0.3 cm2).Accumu- lation of ciliates is an exponential function of time as shown by the semi-log presentation in Fig. 6, and time to reach equilibrium distribution is proportional to the available area of the chamber (Fig. 7). The attraction is obviously chemically induced because ciliates readily accumulate around small capillary tubes filled wlth either peptone solution or bacteria suspension only Fig. 7. Strombidium sulcaturn. Time to reach 90% of the allowing diffusion of soluble substances. equilibrium abundance in a patch of bateria as function of When ciliates swim in gradients of peptone or bac- chamber length terial exudates 2 distinct swimming patterns can be distinguished. Starving ciliates (Fig. 1, right) or ciliates in areas with low bacterial concentrations (Fig. 8, left) swim in helices with occasional tumbles so that transportation rate depends on the steepness of the helix. In contrast, well fed or feeding ciliates (Figs. 1, left and 8, . * M-...... - right) swim in closed circles and more or less stay in one spot. When swimming in circles a ciliary reversal occurs about once in every revolution so the cells swim in a new orbit with a slightly different plane. This kinetic response to the presence of bacteria will cause ciliates to accumulate in a food. When Strombidium Fig. 8. Strombidium sulcatum. Paths plotted from video sulcatum moves into increasing concentrations of bac- recordings of swimming ciliates. Time interval: 40 ms. The 2 teria, tangential velocity (swimming speed through the paths to the left are characteristic of starving cells, the 2 paths water) drops by about 40 %, but due to the change in to the nght for feeding cells. Tumbles are indicated by stars swimming pattern from a helix to a nearly closed circle transportation rate is reduced by more than 95 % (Fig. The swimming pattern of Strombidium sulcatum and 9). Fig. 10 shows the modest range of tangential vel- many other ciliates is characterized by a helical path. ocities found and the corresponding great span of The helical movement can be thought of as a result transportation rates. of 2 superimposed rotations. Due to a propulsion Mar Ecol. Prog. Ser. 48: 1-15, 1988

a.. 0 3456789 log bacteria rnl-'

Fig. 9. Strombidium sulcatum. Tangential velocity (0) and transportation rate (0) as function of bacterial concentration Fig. 11. Schematic representation of a helical motion caused by 2 superimposed rotations. Due to propulsive asymmetry a body moves in a circle with rotation rate oo and radius ro. When a rotat~onaround the long axis, W,, is added a hel~cal path results where r, is the radius of the helix, L and s are distance along the z-axis and the length of the heIix during 1 period [T = 2 nl(ol +oo)] respectively, and 0 is the pitch angle of the helix

clockwise when viewing the anterior pole and close observations reveal that axial rotation ranges between 0.5 and 3.5 S-' in ciliates swimming in helices and is zero for ciliates swimming in circles. As discussed below this axial rotation must be caused by the membranelles generating water currents with a tangential component thus inducing a torque. When these 2 rotations are combined the swimming path will possibly be helical and the steepness of the helix will increase with the ratio wl/wo. The characteristics of the helix and its dependence on the 2 rotational components can be summarized in a mathematical description of transportation rate v, as a function of tangential velocity oo ro and the ratio wl/wo:

0 .l 0.5 1 2 tangential velocity, mm s-l A similar expression was used by Blake & Sleigh Fig. 10. Strombidium sulcatum. Transportation rate as func- (1974). That this is so can be seen from the following tion of tangential velocity considerations. The tangential velocity is always ro wo = s/T (Fig. 11). The period, T = 2n/(wl + wo) since the asymmetry, S. sulcatum tends to turn and thls leads to a same side of the cell is always directed toward the axis, circular path with a characteristic angular velocity, oo z; so s = 2xro wO/(wl + w0) From Fig. 11 it is seen that (Fig. 11). An asymmetry in the ciliary propulsion may L/2n rl = tol/wo and that s2 = (ZX)~~+ L2. Finally the be inferred from the morphology of S. sulcatum (Fig. 1) transportation rate, v = WT. Solving these equations where membranelles close to the oral cavity are small for v leads to Eq. (1). or absent. This is supported by the fact that, whether A graphical representation (Fig. 12) shows that when swimming in a circular or a helical path, the oral side w, goes from 0 to infinity transportation rate increases always faces the axis of the movement. The other from 0 and approaches tangential velocity asymptoti- rotational component is a rotation around the long axis cally. When U), is 0 this corresponds to swimming in a of the cell with an angular velocity, wl; it is always closed circle and transportation rate is accordingly 0 Fenchel & Jonsson: Function a1 biology of an oligotrich ciliate 9

not able to offer an explanation of the motility pattern in hydrodynamical terms. The dramatic reduction in transportation rate in Strombidium sulcatum leading to the accumulation of ciliates in bacterial patches is mainly caused by a reduction in transportation rate. Somehow ciliates have the ability to modulate axial rotation suggesting a change in the action of the membranelles. In most previously studied ciliates membranelles move water parallel to thelr plane and perpendicular to the beat of the individual cilia (Machemer 1974, Fenchel 1986). Observations on a hypotnch ciliate, Euplotes moebiusi, revealed that adjacent membranelles within each membranelle beat about % cycle out of phase (Fenchel 1986). A metachronal wave moves along the cilia within each membranelle which draws water from the peristome and outwards, so that each pair of membranelles acts like a peristaltic pump. The membranelles show a symmetric beat cycle where the indis- Fig. 12. Dimensionless presentation of transportation rate (v). tinguishable effective and recovery stroke both contri- period (7,transportation rate per T (L)and radius of helix (r,) bute to water flow explaining the observed continuous as a function of wl/oo flow. In the heterotrich Stentor, however, Sleigh & hello (1972) found that the membranelles move water and when W, goes to infinity the swimming path at an acute angle to the effective stroke of individual approaches a straight Line. Thus, by modulating the cilia. The detailed function of membranelles in oligo- ratio oo/ol a ciliate may in principle reduce the trans- trich ciliates has not previously been described and in portation rate from near tangential velocity to 0 without an attempt to learn more about their action we video changing tangential velocity. In Strornbidiurn sulcatum recorded swimming ciliates in strobe-light obtaining a this is apparently achieved by changing the rate of series of still pictures with very short exposure times rotation around the long axis wl. Eq. (1) does not (Figs. 14a to c). From these pictures the possible func- describe the swimming paths perfectly since there is tion may be inferred. some reduction in tangential velocity accompanying a The membranelles show a symmetric beat cycle (fre- reduction in transportation rate (Figs. 9 and 10). How- quency ca 35 Hz) with no observed difference between ever, when swimming pattern is analyzed from video effective and recovery strokes (Fig. 14b). Within a recordings there is a reasonable agreement with the membranelle, cilia beat metachronically resulting in a predictions of Eq. (1) (Fig. 13). It must be stressed that wave that travels from the anterior to the posterior edge, evident as changes of the membranelle plane. There is also a metachronic wave pattern among membranelles with about 7 membranelles in each cycle. The function of the membranelles thus seems very similar to that found in Euplotes moebiusi (Fenchel 1986) suggesting a similar mechanism for water transport where the ciliary waves within a membranelle move water in the direction of the metachronic wave. The findings that water transport must be parallel to the membranelles imply that membranelles should be directed radially to prevent a torque which would cause the cell to rotate around its long axis (Fig. 15). On Fig. 13. Strombidium sulcatum. Transportation rate as func- the other hand, to induce rotation membranelles should tion of axial rotation rate (legends as in Fig. 11) change into a turbine configuration generating water currents with a tangential velocity component. Exami- the above is a formal description of what we believe nation of video recordings in strobe-light of Strom- goes on rather than a physical explanation. To achieve bidium sulcatum swimming in circles and helices this it must be shown that the only net force acting on reveal differences in membranelle position. When the cell is directed towards the axis of the helix. We are swimming in circles the membranelles remain in the 10 Mar Ecol. Prog. Ser. 48: 1-15, 1988

Fig. 14. Strombidum sulcatum. Vid- eo-recorded still pictures of chates exposed in strobe-light. (a) Ciliate swimming in c~rcles;(b) clliate swimming in a helix; (c) ciliate swimming in a helix viewed from the side. Arrows show the ciliary beat direction at the anterior rim of the membranelles

same plane during the beat cycle and the membranelle bases seem close to the anterior pole making the rnembranelles appear like spokes in a wheel (Fig. 14a). When swimming in helices the membranelle bases are more distant from the peristome with a tendency to a more tangential position (Fig. 14b). In addition, at each moment the distal parts of several mernbranelles are out of focus and have an apparent shorter length when viewed from above (Fig. 14b),indicating that beating of the membranelles has an up-down component. This is also apparent when rnembranelles are viewed from the side (Fig. 14c) These observations of mernbranelle position and beating planes are consistent with a change from radially orientated membranelles in circling ciliates. The beating pattern of membranelles assuming this change in orientation and based on the observed rnetachronic patterns is shown diagramati- rally in Fig. 15 and explains beating patterns observed It shows how membranelles will have a more up-down beating component when they are tangentially oriented compared to the radially set mernbranelles. This is a consequence of the insertion of the membranelles on the curved cell surface so that a tangential displacement will add an up-down component to the bending plane of the membranelles. How Strombidium sulcatum achieves this change of Fig. 15. Simulation of rnernbranelle beating patterns as seen mernbranelle orientation is unknown. A hypothesis is from the side assuming (above) radially oriented mernbran- that the shape of the cell is slightly changed through elles and (below) membranelles oriented with a tangential component. Arrows show the direction of the thrust acting on the action of contractile elements as known from other the water ciliates. A sphincter-like contraction of the peristome Fenchel & Jonsson: Function ~albiology of an oligotrich ciliate 11

area might impose a strain on the membranelle bases swim in circles the transportation rate is less than 50 pm causing them to change orientation. Indications that S-', so giving a low value of diffusivity; with a sinlung contractions of the peristome area in S. sulcatum are velocity (22 +2 pm S-') they are effectively drawn possible come from observations on ciliary reversals downwards. Upon starvation ciliates begin to swim in either in fixed (Fig. la) or in living ciliates (Fig. 16), helices with transportation rates exceeding 400 pm S-' and so the cells tend to diffuse up in the water column. On the basis of transportation and sinking rates the theoretical model suggested by Roberts (1970) can explain some of the observed distributions. In some experiments with l50 mm deep water columns (Fig. 17) distributions were much more skewed towards the surface and this cannot result from random diffusion and gravitational sinking alone. Similar accumulations Fig. 16. Strombidium sulcatum. Tumbling behaviour by ciliary towards the surface in the absence of other stimuli are reversal (based on video recording). Time interval between known from other ciliate species in spite of the absence drawings is 40 ms of apparent gravity sensors. This upward drift has been interpreted as being caused by the interaction between where the peristome is dramatically contracted and the sinking velocity, motihty and cell asymmetry (Roberts membranelles directed anteriorly. 1970). A front-rear asymmetry of e.g. shape will cause A different type of behavioural pattern related to the the centre of drag and the centre of gravity to be change in transportation rate was found when the separated resulting in a torque which tends to align the vertical distribution of Strombidium sulcatum was front-rear axis vertically. Some observations of S. sul- studied. Well-fed ciliates always accumulate at the caturn indicate that a hydrodynamical asymmetry is bottom but as the food resource becomes depleted present. If fixed cells of S. sulcatum are allowed to sink ciliates begin to move up in the water column (Fig. 17). freely they rapidly (20 t 2" S-') reorientate until the rnembranelle zone points upwards. In addition, by treating starved S. sulcatu~nwith NiN03 or with artifi- cial, calcium-free seawater they swim in steep helices o starved almost without tumbles. This results in a pronounced 0 E 6-8 min after feeding accumulation at the surface and video recordings reveal that many ciliates show swimming paths curving upwards. From the analysed swimming paths reorientation rates can be calculated which are about half the cn -E t rates (13 k 2" S-') found for fixed ciliates. The cause of E 4: O\ this reorientation is obscure, but it may be an :--Q- :--Q- 8-0 0 asymmetry in shape where at least in starved ciliates , , ,?>O# -m-, , , , , , 0 1 the membranelle zone may shift the centre of drag to 1 10 100 number of cells the anterior and thus induce reorientation with the membranelles upwards. The question of vertical distribution of Strombidium sulcatum is further complicated by the amazing behaviour shown by well-fed ciliates when resus- Q) top pended from the bottom. Immediately, ciliates dart Q)a towards the bottom at velocities exceeding 1500 pm S-' 2 4 6 and in 30s most ciliates are back to the bottom. If days of starvation ciliates had to rely on sinking velocity this would not be possible. This behaviour is difficult to explain, but one Fig. 17. Strombidium sulcatum. Above: vertical distribution in clue may be the observation that it was only found in a spectrophotometer cell for starved and well-fed ciliates. Below: vertical distribution (measured as % cells at top and well-fed individuals with large cell size. An increase in bottom) in a 150 mm deep water-column as function of starva- cell volume in S. sulcatum is accompanied by a change tion of the cell shape since growth is localized to the anterior part of the cell with almost no change of the This may in part be explained as a balance between posterior end and the membranelle zone (Fig. 1). One random motility and sinking velocity of the ciliates as suggestion is that the large anterior end will tend to described by Roberts (1970). When well-fed ciliates sink faster than the small posterior end following 12 Mar. Ecol. Prog. Ser. 48: 1-15, 1988

Stoke's law and that this will cause reonentation with Exploitation of patches the anterior end downwards. This effect may dominate over an opposite reorientation tendency in smaller The behaviour of Strombidum sulcatum reflects an cells. Experiments with ciliates swimming in seawater adaptation to life in an environment where food is with increasing density show that the tendency to swim patchily distributed. The basic behavioural response to downwards is reduced and eventually ceases in a high- prey density is one of chemokinesis, that is, a modula- density medium (Table 1). The experiments show that tion of motility as a function of the concentration of some chemical cue (Fig. 18). When a population of S.

Table 1. Stron~bidiumsulcatum. Average absolute swimming angle (+ SD) to the vertical as function of medium density. Medium with the lowest density was made from seawater and the higher densities from mixtures of seawater and deuteriumoxide

Density (g cm-3) Swimming angle (") I I . ...*

.=9.. \ .- ..,g: bacteria *. .* Sample size . ..'

average absolute swimming angle to the vertical increases from 7 + 7" in seawater, indicating a high directionality, to about 45 f 29" in high density medium, which is expected for random motility. This reduction of directed movement is expected if the ability to swim downwards is caused by reorientation induced by a separation of the centre of gravity and centre of drag since raising the density reduces the effect of gravity and thence the reorientation torque. Although some questions remain concerning the verti- Fig. 18. Strombidium sulcatum. Above: swimming paths plot- cal distributionof s. sulcatum it can be concluded that ted from video recordings of 2 individuals entering a patch of when ciliates are growing and have a large cell size bacteria (time interval: 80 ms). Below: swimming path of a cell in the vicinity of a peptone agar block (time interval 200 ms) they concentrate close to the bottom. The equilibrium distribution can be explained by the low motility and gravitational sinking, but the rapid return to the equi- sulcatum is allowed to spread in an environment with librium when ciliates are resuspended requires a patches of food the process can be viewed as a diffusion directed movement downwards. It is suggested that process in which diffusion coefficients differ in- and downward swimming is caused by hydrodynamical outside patches. After some time a dynamic equi- reorientation. When ciliates have exhausted their food librium is reached where the concentrations in- and resource they start to swim in helices and the motility outside patches are constant. The equilibrium concen- increases. This accounts in part for the upward move- tration of cells is everywhere inversely proportional to ment, but in addition starving chates, whlch have a their local mot~lity,thus different shape, tend to reorientate with the anterior end upwards giving the ciliates a directed movement towards the surface. Thus it seems to be a delicate Where C,, = cell concentration in the patch; C, = cell balance of swimming pattern, slnking velocity and concentration outside the patch; D, and D, = local nutritional status that determines the vertical distribu- motilities (Okubo 1980). A time-dependent solution is tion of S. sulcatum. It may well be that most of these likely to depend on details of the system such as the features are nonadaptive, but the general increase of geometry of the patch. However, from dimensional diffusion of cells upon starvation which affects both considerations it is reasonable to assume that the time horizontal and vertical distribution suggests an adap- taken to reach equilibrium is proportional to the square tive dispersal of cells to new patches. of a length measure of the (closed) system and that Fenchel & Jonsson: Functional biology of an oligotnch ciliate 13

deviations from the equilibrium decay exponentially 0.2 with time at a rate which is proportional to the motility. To test if the behaviour of S. sulcatum accumulating in a bacterial patch (Fig. 6a) conforms to a distribution process as described above, estimates of the diffusion coefficients of the 2 modes of swimming are required. - 0.1 Estimates could be obtained in 2 ways. From analysis of swimming tracks transportation rate (v) and time 2S between tumbles (t) can be directly measured and the 5 diffusion coefficient (D) in a 2 dimensional environment approximated as: 0 1 0:, 1 2 3 4 D = 9,.~/4.(Okubo 1980) (3) distance from agar, mm Using this expression we found motilities of 0.0012 and 0.21 mm2 S-' for feeding and sta~ngcells, respec- Fig. 19. Strombidium sulcatum. Tumbling frequency as func- tively or a ratio D,/Df of about 175 (where D, and Df tion of distance from a peptone agar block designate motilities of starved and feeding cells, respectively). Another method is to use the differences famous example of a transient response is the avoidance in vertical distribution between well-fed circling reaction which prevents ciliates from entering ciliates and starving ciliates swimming in helices. unfavourable areas (Jennings 1906). When a ciliate Assuming no hydrodynamical reorientation of cells the swims up a gradient of some harmful substance it number of ciliates (X) will be an exponential function tumbles when some characteristic rate of change is (Fenchel & Finlay 1984) of height from bottom (h), exceeded and if the ciliate after tumbling swims down sinking velocity (U),and diffusion coefficient (D) given the gradient tumbling is suppressed and it will escape by: from the area. Note that a kinetic response in a similar X(h) = e-hU/D (4) situation could be fatal where a kinetic induction of Accordingly the slopes of the semi-logarithmic verti- tumbling to a raised concentration might lead to a cal distributions will equal ulD which experimentally trapping of the ciliate in a harmful patch. (Fig. 17) were found to be about -0.025 and -1.45 for The transient response shown by Strombidium sul- well-fed and starving ciliates respectively. With a catum..when swimming in a peptone gradient is similar measured sinking velocity of 22 + 2 pm S-' the result- to the avoidance reaction, but here tumbling is trig- ing diffusion coefficients are 0.88 and 0.015 mm2 s-' gered when ciliates swim down a sufficiently steep giving a ratio of about 60. grahent of an attractant (Fig. 18). Studies of many The experiments on cell accumulation such as the individuals showed that the induction of tumbling one shown in Fig. 6 gave somewhat variable results, occurs at about 2 mm from the agar block presumably but in general the concentration of ciliates in the bac- indicating the level of concentration for which the terial patch was some 60-fold higher inside the patch. ciliates are most sensitive to changes (Fig. 19). While the above estimates of cell motility are inaccu- Whenever a ciliate approaches the outskirts of a food rate for a number of reasons the results suggest that the patch it wdl tumble and thereby remain in the vicinity accumulation can be explained as a kinetic response. of the food patch centre and when the ciliate eventually However, careful studies of Strombidium sulcatum encounters the patch centre a kinetic response is swimming in an environment with a block of peptone induced which reduces motility. In a way the transient agar revealed another behavioural response in addition response makes a patch appear larger and the to the described chemokinesis (Fig. 18). This behaviour encounter rate of new patches will be enhanced. It is is known as a transient response (see Fenchel 1987 and more likely to play a role in keeping the cells in close Fenchel & Finlay 1986b for a discussion on terminol- vicinity to a prey patch as a 'fine tuning' of the kinetic ogy) and consists of tumbles or suppression of tumbles response since sufficiently steep chemical gradients in response to the rate of change of some environ- are not Likely to extend very far from the patch. mental cue perceived by the ciliate during one run (the It may be concluded that the attraction of Strom- length moved between tumbles). The response is called bidium sulcatum to food patches can largely be transient since it is triggered only where the gradient is explained as a chemokinetic response. Through a sufficiently steep while on either side of the gradient reduction in axial rotation, motility is reduced when a the response is absent. In this way it differs from a certain concentration of food is encountered and upon kinetic response which results from different motilities food depletion motility is again increased and the in different, homogeneous environmental conditions. A ciliates spread to new patches. The shlft in motility Mar. Ecol. Prog. Ser. 48: 1-15, 1988

occurred at a threshold of 106 to lo7 bacteria ml-' (Fig. grow S. sulcatum on microalgae and at least the plank- g), a higher concentration than found in free seawater. tonic S. vestitum can grow on bacteria and hetero- This makes sense since no shift in motility would occur trophic flagellates (pers. obs.) indicahng that they if the concentration threshold was similar to or lower could survive and reproduce in either habitat. One than found between food patches. Note that a serious difference in planktonic Strombidium species is their behavioural constraint is the inabiltty to directly sense seemingly weak ability to respond to food gradients the direction of a chemical gradient, which S. sulcatum (Jonsson unpubl.). This probably indicates that their shares with most protozoa. prey is more or less homogeneously distributed in their An interesting question is on what length scale the pelagic habitats, at least on the scale of ciliates. How- behaviour of Strombidium sulcatum is appropriate. ever, selection of habitat may have a proximal cause How far can ciliates travel in search for new food not coupled to prey availability. Since the pelagic and patches? The time to find a bacterial patch was found to benthic habibtats are separated vertically, habitat increase with the square of the length of the chamber selection could be achieved by behaviour that causes (Fig. 7). If ciliates which leave an exploited patch are different species to swim either upwards or down- assumed to diffuse with a diffusion coefficient D, the wards. This is supported by the studies of the vertical time, t, to move a certain distance, 1, can be approxi- distribution of S. sulcatum discussed above together mated as: with observations showing that many planktonic species have a net upward swimming movement probably caused by hydrodynamical reorientation of In the growth experiments it was found that S. sul- asymmetric cells (Jonsson unpubl.). catum could survive starvation for maximally about The behaviour of Strombidium sulcatum shows how 100 h (Fig. 2). Extrapolation of the data shown in Fig. 7 a few simple mechanisms suffice for exploiting a shows that a ciliate could on average move ca 0.5m heterogeneously distributed food resource. Many pro- before succumbing to starvation. This indicates the tozoa swim in a helix similar to S. sulcatum and the scale of patchiness which S. sulcatum is adapted to ability to modulate transportation rate by changing exploit. axial rotation may be widespread among species living A kinetic response implies that motility is reduced on patchily distributed food, or wherever a kinetic when conditions are favourable; however, the mem- response to environmental change is desirable without branelles serve as a filtering apparatus for food parti- cutting swimming speed. cles as well as for locomotion. The only 2 ways to reduce the motility to zero when a food patch has been Acknowledgements. We are indebted to Jeanne Johansen encountered is therefore either to attach to some and Ilse Duun for assistance with experimental work. We also thank Dr Fereidoun Rassoulzadegan who provided the culture immobile surface or to continously move in a circle. of Strombidium sulcatum isolated from the Mediterranean Strombidium sulcatum does not have the ability to Sea. The study was supported by the Danish Natural Science attach and the transformation from helical swimming to Research Council (Grants Nos. 11-6089, 11-6480, 11 -4688), a closed circle is the best solution. In reality the ciliates the Carlsberg Foundation, the Swedish Natural Science do not swim in a constant circle but due to tumbling Research Council through contract B-DT 1860-107 and by a Grant from the Nordic Council for Marine Biology. about once every revolution the circular path is slightly rotated and translated resulting in the observed slow diffusion rates around 0.01 mm2 S-' (Fig. 8). These LITERATURE CITED small recurrent perturbations of the circular path are probably an adaption to avoid refiltering water already Andersen. P.. Ssrensen, H. M. (1986). Population dynamics cleared for prey. It is interesting to note that some and trophic coupling in pelagic microorganisms in eut- rophic coastal waters. lMar Ecol. Prog. Ser. 33: 99-109 benthic Strombidium species seem to have adopted the Blake, J. R , Sleigh, M A. (1974) Mechanics of ciliary locomo- attachment strategy (Kahl 1932, Faure-Fremiet 1950, tion. Blol. Rev. 49: 85-125 1969).These species are reported to possess either long Borror, A C. (1965).New and little-known tidal marsh clliates. thigmotactic membranelles or mucous threads which Trans. Am. Micros. Soc. 84: 550-565 allow for temporary attachment to surfaces and Faure-Fremiet, E. (1924). Contribution a la connaissance des infusoires planktoniques. Bull. Biol. Fr Belg. 6 (Suppl.): although not studied it is likely that attachment and 1-171 detachment are controlled by prey density. Faure-Fremiet, E. (1950). Ecologie des Cilies psam.mophiles Most known Strombidium species are planktonic, littoraux. Bull. Biol. Fr. Belg. 84: 35-75 but very similar in morphology to S. sulcatum and other Faure-Frerniet, E. (1969).Remarques sur la systematique des Cilies Oligotrichida. Protistolog~ca5: 345-352 benthjc species (Kahl 1932). Why are some species Faure-Fremiet, E,Ganier, M. M -Cl. (1970).Structure fine du found in the pelagic and others in benthic habitats? Strombidium sulcatum Cl. et L. (Ciliata Oligotrichida). Their particle selection is similar and it is possible to Protistologica 6: 203-223 Fenchel & Jonsson: Functional biology of an oligotrich ciliate 15

Fenchel, T. (1982). Ecology of heterotrophic microflagellates. Tontonia appendiculariformis (Ciliophora. Oligotrichina). 11. Bioenergetics and growth. Mar Ecol. Prog. Ser. 8: BioSystems 19: 137-158 225-231 Leergaard, C. (1915) Untersuchungen iiber einige Plankton- Fenchel, T. (1986). Protozoan filter feeding Prog. Protistol. l. clliaten des Meeres. Nytt Mag. Naturv. 53: 1-37 65-1 13 Levandowsky, M., Hauser, D. C. R. (1978). Chemosensory Fenchel, T (1987). Ecology of Protozoa. Science Tech/ responses of swimming algae and protozoa. Int. Rev. Cytol. Springer Verlag. Madison 53. 145-210 Fenchel, T., Finlay, B. J. (1983). Respiration rates in hetero- Machemer, H. (1974). Ciliary activity and metachronism in trophic, free-living Protozoa. Microb. Ecol. 9: 99-122 Protozoa. In: Sleigh, M. A (ed) Cilia and flagella. Fenchel, T., Finlay, B J (1984). Geotaxis in the ciliated proto- Academic Press, London, p. 199-286 zoon Loxodes. J. exp. Biol. 110: 17-33 McManus, G. B., Fuhrman, J. A. (1986). Bacterivory in sea- Fenchel, T., Finlay, B. J. (1986a). The structure and funchon of water with the use of inert fluorescent particles. Limnol. Miiller vesicles in loxodid ciliates. J. Protozool. 33: 69-76 Oceanogr. 31: 420426 Fenchel, T.,Finlay, B. J. (1986b). Photobehavior of the ciliated Naitoh, Y., Eckert, R. (1974). The control of ciliary activity in protozoon Loxodes: taxic, transient, and kinetic responses Protozoa. In: Sleigh, M. A. (ed.) Cilia and flagella. in the presence and absence of oxygen. J. Protozool. 33: Academic Press, London, p. 305-352 139-145 Okubo, A. (1980). Diffusion and ecological problems: Finlay, B. J., Fenchel, T. (1986). Photosensihvity in the clliated mathematical models. Springer-Verlag, Berlin protozoan Loxodes: pigment granules, absorption and Reisser, W., Hader, D.-P. (1984). Role of endosymbiotic algae action spectra, blue light perception, and ecological sig- in photokinesis and photophobic responses of ciliates. nificance. J. Protozool. 33: 534-542 Photochem. Photobiol. 39: 673-678 Finlay, B. J., Berninger, U,-G., Stewart, J. L., Hindle, R. M,, Rivier, A., Brownlee, D C., Sheldon, R.W., Rassoulzadegan, F. Davison, W. (1987). Some factors controlling the distribu- (1985). Growth of microzooplankton: a comparative study tion of two pond-dwelhng ciliates with algal symbionts of bacterivorous zooflagellates and ciliates. Mar. Microb. (Frontonia vernalis and Euplotes daidaleos). J. Protozool. Food Webs 1:51-60 34: 349-356 Roberts. A. M. (1970). Geotaxis in motile micro-organisms. J. Heinbokel. J. F. (1978). Studies on the functional role of exp. Biol. 53: 687-699 tintinnids in the Southern California Bight. I. Grazing and Sleigh, M. A., Aiello, E. (1972). The movement of water by growth rates in laboratory cultures. Mar Biol. 47: cilia Acta Protozool. 9. 265-277 1977-189 Song, P. S., Hader, D -P,,Poff, K. L. (1980). Phototacbc orienta- Inoue. S. (1986). Video microscopy. Plenum Press, New York tion by the ciliate Stentor coeruleus. Photochem. Photobiol. Jennings, H. S. (1906). Behavior of the lower organisms. Col- 32: 781-786 umbia Univ. Press, New York Stoecker. D. K., Michaels. A. E., Davis, L. H. (1987). Large Jonsson, P. R. (1986). Particle size selection, feeding rates and proportion of marine planktonic ciliates found to contain growth dynamics of marine planktonic oligotrichous functional chloroplasts. Nature, Lond 326: 790-792 ciliates ( Ciliophora: Oligotrichina). Mar. Ecol. Prog. Ser. Van Houten, J., Hansma, H., Kung, C. (1975). Two quantita- 33: 265-277 tive assays for chemotaxis in Paramecium J. comp. Physiol. Jonsson, P. R. (1987). Photosynthetic assimilation of inorganic 104: 211-223 carbon in marine oligotrich ciliates (Ciliophora, Oligo- Van Houten, J., Hauser, D. C. R., Levandowsky, M. (1981). In: trichina). Mar. Microb. Food Webs 2: 55-68 Levandowsky, M., Hutner, S. H. (eds.) Biochemistry and Kahl, A. (1932). Urtiere oder Protozoa I Wimperntiere oder physiology of protozoa, Vol 4. Academic Press, New York, Ciliata (Infusoria). 3. Spirotricha. In: Dahl, F. (ed ) Die p 67-124 Tienvelt Deutschlands und der angrenzenden Meeres- Zobell. C. E. (1941). Studles on marine bacteria, I. The cultural teile. Gustav Fischer Verlag, Jena, p. 399-650 requirements of heterotrophic aerobes. J. mar. Res. 4: Laval-Peuto, M., Febvre, M. (1986). On plastid symbiosis in 42-75

This article was submitted to the editor: it was accepted for printlng on June 9, 1988